1. Field of the Invention
The present invention relates to a connectionless communications system for transmitting data at a high speed, to a method of testing the system, and to an intra-station control system of a switching station for transmitting data at a high speed.
2. Description of the Related Art
Recently, high-performance information processing devices such as work stations, personal computers, etc. have been developed to perform a distribution process in which a number of information processing devices are interconnected through a high speed local area network (LAN). The network connecting such LANs should also be provided with high speed processing capabilities.
One of the services to realize the above described high speed data communications is a switched multi-megabit data service (SMDS). The SMDS is a connectionless data switching service based on the transfer speed of 1.5 Mbps and 45 Mbps.
An asynchronous transfer mode (ATM) system is well known as a method of realizing a broadband ISDN, and the SMDS can be provided through an ATM network. In this case, an SMDS processing server (SMDS message handler) is supplied for a predetermined ATM switch, and a permanent virtual circuit or a permanent virtual channel (PVC) connects an SMDS subscriber with the SMDS processing server accommodating the SMDS subscriber. The connectionless data output from the SMDS subscriber is transferred to the SMDS processing server to perform a routing process, etc. at the server.
The connectionless data normally refers to a variable packet (data frame). However, since the above described PVC is a path to be established in a network, the connectionless data is transferred after being converted (decomposed) into an ATM cell format before it is input to the ATM switch. The cell is a 53-byte structure consisting of a 48-byte payload and a 5-byte header.
The ATM cell format data is temporarily structured as the layer-3 protocol data unit (L3-PDU) or in a data format of a higher-level layer in the SMDS processing server as shown in FIG. 897 to analyze routing information, etc. according to a destination address DA, a source address SA, etc. stored in the L3-PDU. Then, the data is decomposed again into cells to route the data according to the analyzed information.
As described above, the conventional SMDS is limited in its speed because input cells are structured in a higher level layer data format (for example, in L3-PDUs) when the SMDS processing server performs a routing process through software of a microcomputer program, etc. Additionally, such processes as a data copying process performed when a group address is specified as a destination address DA, a traffic smoothing process, an action against no reception of an end-of-message cell (EOM: a cell storing the last portion of data when an L3-PDU is decomposed into a plurality of cells) have been processed through software by microcomputers, etc.
Thus, the conventional SMDS has been limited in its process speed because the processes in the SMDS processing server are performed through various software. Therefore, when connectionless communications data is transmitted using an SMDS, the operations of the transmission line and switch are sped up with the SMDS processing server processes interfering as a bottleneck, thereby preventing an actual high-speed process from being successfully realized. Furthermore, when the above described structuring process in the SMDS processing server, all cells forming each L3-PDU should be temporarily stored. Therefore, the necessary buffer capacity undesirably becomes very large.
In the SMDS, protocol performance is monitored when a service is offered as follows. That is, the formats of various parameters are checked in the data, and counted is the data which has been rejected by the check (the data which cannot be recognized as valid). A predetermined specific type of check is followed by a counting process performed on the rejected data based on a predetermined algorithm. If the resultant value exceeds a predetermined threshold, then output is a threshold crossing alert (TCA) indicating that the threshold is exceeded. Furthermore, an error log is collected each time data is rejected.
The following parameters are collected in the error log.
(1) Destination address DA PA1 (2) Source address SA PA1 (3) SNI number (subscriber network interface No.) PA1 (4) Error type PA1 1. Copying a cell PA1 2. Reassigning a VPI/VCI
In the PVC between the user (subscriber) and the SMDS processing server,
In the PVC between the user (subscriber) and an SMDS processing server, data is transferred in the cell format as described above (actually, the data is transmitted in the ATM cell format and processed in the L2-PDU in the SMDS processing server. The ATM cell and L2-PDU are based on the 53-byte configuration and simply referred to as cells. However, since the above described error log collection is mostly related to the layer 3, the data is received in the cell format and then reassembled into the L3-PDU in the SMDS processing server.
As described above, input cells are reassembled in the data format of the higher order layer (for example, L3-PDU) in the conventional SMDS. This prevents the processes from being performed at a high speed in the SMDS.
The above described services are based on the high reliability of the physical quality of the transmission lines forming the network. Therefore, it is important to test and evaluate the transmission quality of the network.
The test and evaluation of the transmission lines are activated from the OS center (operation center for managing the network) in the connectionless communications service network, and an inter-station loopback test is conducted to confirm the normality of any inter-station link (path between switches). The inter-station loopback test is described below by referring to FIG. 898. In this embodiment, the test is conducted to check the link between SW station 3 and SW station 6.
The test is started by issuing a test connectionless packet transmission request message (test start request) from the OS center 1 to SW station 3. The request message contains an identification information ID indicating terminal SW station 6. SW station 3 generates a test packet with the identification address of terminal SW station 6 set as its destination address DA and the identification address of its home station (SW station 3) set as its source address SA. The test packet is output to terminal SW station 6. In SW stations 4 and 5, test packets are processed as normal packets and transferred to terminal SW station 6. On receipt of the test packet, terminal SW station 6 outputs the packet with its DA and SA inverted. That is, the packet is returned from terminal SW station 6 to SW station 3, and it is reported to the OS center 1 upon re-arrival of the packet at the source SW station 3.
Thus, the OS center 1 checks whether or not the packet is normally transmitted in the network, that is, checks the normality of the transmission line (the link between SW station 3 and terminal SW station 6 in this embodiment). In the procedure, since the source SW station 3 and the terminal SW station 6 mark the time stamp onto the payload field of the packet, the OS center 1 is informed of the transmission time of packets according to the information.
However, in the above described test method, the information obtained by the test is to be provided for the OS center (operation center), and no method has been provided for the subscriber (terminal unit 2 in FIG. 898) to be autonomously informed of the transmission quality in the network (transmission delay time, etc.). Therefore, if a packet is not normally transmitted from a source subscriber to a destination subscriber, the subscribers cannot detect in which the factor of the fault resides, the subscriber terminal unit or the network transmission line. Thus, the OS center is invoked to recover from the fault, thereby requiring much time and cost.
FIG. 899 shows an embodiment of the SMDS. In FIG. 899, the SMDS support module analyzes a destination address DA and makes various checks. An SMDS support module S accommodates a plurality of source SMDS subscribers (a) and (b) to analyze a DA and make various checks. The SMDS support module R accommodates a plurality of destination SMDS subscribers (x) and (y) to make various checks. The modules comprising these S and R correspond to the above described SMDS processing server (SMDS message handler).
Each of the source SMDS subscribers (a) and (b) is connected to the SMDS support module S through the PVCs 1 and 2. The SMDS support module S is connected to the SMDS support module R through the PVC 3. The SMDS support module R is connected to each of the destination SMDS subscribers (x) and (y) through the PVC 4 and 5.
If the SW shown in FIG. 899 comprises an ATM switch, the connectionless data (SMDS message) output from the source SMDS subscribers (a) and (b) is converted into the cell format in the interface not shown in FIG. 899. The cell is transferred to the SMDS support module S by assigning to the header of the cell a specific VPI/VCI specifying the SMDS support module (VPI/VCI specifying the PVC 1 and 2) as its destination. In the transfer between the SMDS support modules S and R, the VPI/VCI value indicating the PVC 3 is assigned and output. The cell transferred from the SMDS support module R to the destination SMDS subscribers (x) and (y) with a specific VPI/VCI value indicating the PVCs 4 and 5 is output from the SMDS support module R, and arrives at the destination SMDS subscribers (x) and (y). Each of the PVCs is established at the system initialization.
Since the numbers of the source and destination SMDS subscribers accommodated in the SMDS support modules S and R are limited, a plurality of SMDS support modules are provided if a single SW station accommodates SMDS subscribers in excess of the maximum number. FIG. 900 shows an example of this. In this case, each connection is made by the PVC. FIG. 900 shows an example that SMDS subscribers (a), (b), (x), and (y) are accommodated in the SMDS support module 1 and SMDS subscribers (c), (d), (v), and (w) are accommodated in the SMDS support module 2. The PVC also connects SMDS support module 1 to SMDS support module 2.
As described above, the data transfer path is set at the system initialization in the SMDS. If the source SMDS subscribers (a) and (b) output SMDS messages, the messages are led to the SMDS support module S through the PVCs 1 and 2, and transferred to the destination SMDS subscribers (x) and (y) through the PVCs 3, 4, and 5. Therefore, it cannot be verified that the SMDS messages output from the source SMDS subscribers (a) and (b) have arrived at the destination SMDS subscribers (x) and (y) through the PVCs.
If the data cannot be successfully transferred, a complaint is expected from the source SMDS subscribers (a) and (b) or destination SMDS subscribers (x) and (y). The subscriber's complaint should be appropriately verified at the lowest possible cost.
The PVC test and the transmission time test are described above, and the SMDS needs confirming the normality of the transmitted SMDS data. The method of confirming the normality of data includes checking the BS-size of the L3-PDU, length of the L2-PDU, etc.
In the BA-size check, it is confirmed whether or not the value for use in checking the payload length of the L3-PDU (CPCS-PDU) is correct. In the BE-tag (beginning tag and end tag) check, the normality of the L3-PDU data can be confirmed by verifying the matching between the leading and trailing tags of the L3-PDU. In the length check, it is confirmed that the assembling and disassembling between the L3-PDU and L2-PDU are normally performed by verifying the relationship between the valid payload length value of the L2-PDU and the BA-size of the L3-PDU.
When the normality of the L3-PDU is confirmed in the disassembled L2-PDUs, the scale of the circuit becomes undesirably large. Since the BA-size and BE-tag of the L3-PDU and the length of the L2-PDU are checked as being closely related to one another, it is difficult to perform a process for each cell (for each L2-PDU). If the data in the format of the cell input to the SMDS processing server (L2-PDU) is processed after being assembled into the L3-PDU, a high-speed process is prohibited by the software process involved as described above.
When the connectionless communications service is realized in the ATM switch network, a connectionless data processing server (SMDS processing server in the SMDS) is provided to request the server to check the routing process on the connectionless data output from the subscriber terminal unit and to make various checks. FIG. 901 shows an example of the method of realizing such connectionless communications services. The configuration shown in FIG. 901 is the same as that shown in FIG. 899. That is, a PVC 11 is set between the source SMDS subscriber (a) and the connectionless data processing server CLS 2. A PVC 13 is set between the destination SMDS subscriber (x) and the connectionless data processing server CLS 6. These PVCs are set using a call processor CPRs 3 and 7.
In the configuration shown in FIG. 901, the connectionless data processing server CLS 2 accommodating the source subscriber (a) and the connectionless data processing server CLS 6 accommodating the destination subscriber (x) are provided in different switch stations. That is, the connectionless data processing server CLS 2 is provided in the SW station 1, while the connectionless data processing server CLS 6 is provided in the SW station 5. These connectionless data processing servers CLS 2 and 6 are connected to each other by the PVC 12. A large-scale relay switch 4, in which the PVC 12 is provided, has the configuration of relaying switches such as SW 1 or SW 5, or is an ATM interconnection switch (AISW).
When connectionless data is transferred from the source SMDS subscriber (a) to the destination SMDS subscribes (x) with the above described configuration, the data output from the source SMDS subscriber (a) is input to the connectionless data processing server CLS 2 through the PVC 11, and then transferred to the connectionless data processing server CLS 6 through the PVC 12. Then, it is transferred to the destination SMDS subscriber (x) from the connectionless data processing server CLS 6 through the PVC 13. The data is transferred through the PVCs in cell units and routed by the connectionless data processing servers CLS 2 and 6.
In the conventional connectionless communications service, the connectionless data processing server CLS 2 accommodating the source SMDS subscriber (a) is connected to the connectionless data processing server CLS 6 accommodating the destination SMDS subscriber (x) through the PVC 12 as shown in FIG. 901 if these servers are different from each other. The PVC 12 is set such that it passes through the SWs 1 and 5, and the large-scale relay switch 4. Therefore, the band resource for connectionless services should be preliminarily reserved in the switches to manage the services.
In the conventional systems, the band resource for each switch is used even when the connectionless service data is not being transmitted, and the band resource management is complicated.
By contrast, the switches for switching cells such as a B-ISDN (broadband ISDN) switch for providing broadband services, for example, ATM (asynchronous transfer mode) services, an SMDS switch for providing SMDS (switched megabit data service) services, etc. require considerably high performances and functions as compared with the conventional telephone switches or N-IDSN (narrowband ISDN) switches. Therefore, these switches require unique technology for intra-station control.
The prior art technology and the problems are clearly described below.
Described below is the problems related to the intra-station control communications technology for communicating the control information between the intra-station devices such as various transmission line interface device (trunk), etc. and the switch processor.
In controlling the intra-station devices in the conventional switching system, each of the intra-station devices 6 and 7 for operating with an ATM switch 5 is connected through an input control device 4 to a system bus 3 to which a switch processor (CC)1 is connected as shown in FIG. 902 to transfer the control information between the intra-station device and a main storage memory (MM) 2 connected to the CC 1 by the direct memory access (DMA) system.
In this system, however, all the intra-station devices 6 and 7 should be connected to the system bus 3, and the cable should be mounted to connect the intra-station devices 6 and 7 to the system bus 3.
Thus, the farther the intra-station devices 6 and 7 are located from the system bus 3, the longer the cable should be, thereby causing the problem of complicated connection.
Connecting all the intra-station devices 6 and 7 to the system bus 3 causes a conflict for the acquisition of an access right required to access the bus, thereby resulting in the congestion of bus access.
Furthermore, extending the system bus 3 to each of the intra-station devices 6 and 7 lowers the transmission quality, and may generate a transmission error such as a data error and parity error in the DMA procedure which includes no error control procedure.
Described next is the problem related to the technology for communicating control information such as call setting information, etc. between a terminal unit and a control device such as a switch processor.
Controlling a terminal interface device in the ATM switch system, etc. requires communicating control information with a control system device such as a switch processor, etc.
The conventional technology to communicate control information can be the system in which a physical interface is connected to a terminal unit (TERM) 4 connected from the control system device (MPR1 and PRIF2) to the switch (SW) 3 as shown in FIG. 903 as in the case shown in FIG. 902.
Since a physical interface is required for each terminal 4 in this system, the entire system configuration is complicated and the problem occurs that the terminal units 4 cannot easily added.
Described below is the subject related to the technology of testing a switch as an intra-station control system.
In the ATM switch, etc, a test is conducted whether or not a cell transmission highway is faulty by connecting to a highway a test device for sending cells and retrieving and collecting received cells.
In this case, a test cell is transmitted after setting the destination information VPI (virtual path identifier), VCI (virtual channel identifier), cell loopback in the test device, and other LSIs through the test device.
However, such a system requires a complicated configuration of a test device, and takes time in setting a test device.
Described below is the loopback test in the technology of testing switches.
With an increasing use of ATM switches and ATM switch network in which the information of different traffic characteristics such as voice, data, animation, etc. can be combined and switched, a test of confirming the normality of an inter-station path has been required. If a fault occurs between the two stations having a lot of stations existing between the two stations in an actual operation, it is required that faults should be detected and corrected at the earliest possible stage. The loopback test method of an ATM switch network is an effective test method for quickly detecting a fault between the stations.
The ATM switch has just been introduced in the market, and the ATM switch has never been tested between stations. However, the following test method is considered to be an effective inter-station ATM switch network test method based on the conventional electronic switch test method.
According to this method, if a number of stations exist in the ATM switch network, a test device should be provided for each test device.
If there are not sufficient test devices, a test device should be shared among stations for the test.
Furthermore, some stations are not constantly attended by operators and the operators should go to the stations to conduct the test.
Thus, in the above described method, operators are required to go to trouble in conducting an inter-station test.
Described next is the subject related to the technology of measuring the performance in a switch according to the intre-station control system.
The self routing module (SRM) switching method using the ATM is the condition for structuring a broadband ISDN system. However, measuring the performance in the SRM has been a difficult task.
Finally, the subject related to the control of a trailer in the PLCP, which is a physical layer conversion protocol interfaced in the DS3 format, that is, the digital signal level 3 format, is described below as one of the intra-station control system.
In the B-ISDN or SMDS service, the DS3 (digital signal level 3) format is used to realize the service of 44.736 MHz.
FIGS. 904 and 905 show examples of system configurations according to the present invention.
FIG. 904 shows the configuration in which the BISDN terminal unit is connected to the BISDN switch.
FIG. 905 shows the configuration in which the SMDS terminal unit is connected to the SMDS switch. The present invention is related to the transmitting units in the BISDN terminal unit and BISDN switch or the SMDS terminal unit and SMDS switch.
FIG. 906 shows the configuration of the DS 3 multi-frames. The DS 3 frame comprises 85-bit basic frames. The basic frame comprises a 1-bit DS 3 header and an 84-bit DS3 payload. Eight basic frames form a subframe, and seven subframes form a single multi-frame. That is, one multi-frame consists of 56 (8.times.7) basic frames.
The ATM cell of the BISDN is a 53-octet cell, and the L2-PDU (level 2 protocol data unit cell) of the SMDS is a 53-byte cell. That is, they are similar in basic configuration, but different in contents of the header and payload and in value of the HEC and HCS.
FIGS. 907(a) and (b) show the configurations of the ATM cell and L2-PDU cell.
An ATM cell or L2-PDU cell are not directly stored in the payload of the DS3 reference frame, and transmitted through the frame of the PLCP (physical layer convergence protocol).
FIG. 908 shows the configuration of the PLCP multiframe interfaced in the DS3 format.
Each of the ATM cell or L2-PDU cell is stored in a 53-octet PLCP payload in the PLCP frame. The PLCP multiframe is divided into 84-bit segments, and each segment is stored in an 84-octet DS3 payload in the DS3 frame and then transmitted.
The PLCP frame is a multiframe comprising 12 pairs of a 4-byte PLCP header and 53-byte PLCP payload and a trailer. The PLCP header comprises A1 and A2 bytes, POHI, and POH. The trailer length is 13 or nibbles. A nibble is 4 bits and refers to a half byte. The trailer data is 13 or 14 4-bit patterns "1100".
One PLCP multiframe is transmitted at an average of 125 .mu.sec (8 KHz cycle). Variable trailer length defines an average value.
Described below is the trailer. Since the DS3 frame is transmitted at a speed of 44.736 MHz, 5592 bits are transmitted in the 125-.mu.sec period according to the following equation. EQU number of bits=44.736.times.10.sup.6 (bit/sec).times.125.times.10.sup.-6 (sec)=5592 bits [equation 1]
However, the data forming the DS3 frame comprises a 1-bit frame bit data and an 84-bit DS3 payload, the number of bits in the DS3 payload for the period of 125 .mu.sec is 5592.times.84/85=5526.211 . . . as not divisible.
The number of bits in the PLCP multiframe is 57.times.12.times.8+13.times.4=5524 bits when the trailer length is 13 nibbles, and 57.times.12.times.8+14.times.4=5528 bits when the trailer length is 14 nibbles. That is, there is a residue in the DS3 payload in the 125-.mu.sec period when the trailer length is 13 nibbles, and there is a deficiency in the DS3 payload in the 125-.mu.sec period when the trailer length is 13 nibbles.
To transmit PLCP multiframes at an average speed of 125 .mu.sec (8 KHz cycle), the PLCP multiframes are transmitted with their trailer length changed between 13 and 14 nibbles.
A C1-byte cycle staff counter is used to display the trailer length (refer to FIG. 908). FIG. 909 shows the definition related to the cycle staff counter.
As shown in FIG. 908, the C1 byte is cyclically changed on three multiframe cycles. In the first multiframe, C1 refers to FF and the trailer length is 13 nibbles. In the second multiframe, C1 refers to 00.sub.H and the trailer length is 14 nibbles. In the third multiframe, C1 refers to 66.sub.H or 99.sub.H and the trailer length is 13 nibbles for C1=66.sub.H and 14 nibbles for C1=99.sub.H. The trailer length of 13 or 14 nibbles is determined such that the PLCP multiframes are transmitted at an average speed of 125 .mu.sec (8 KHz cycle).
Then, there arises a problem as to what the value of C1 of the third multiframe should be, that is, how to control the trailer. Described below is the conventional method of controlling the trailer.
Assuming that the pattern prefers to 13 nibbles for the third multiframe and the pattern Q refers to 14 nibbles for the third multiframe, the number of nibbles for the trailer changes 13.fwdarw.14.fwdarw.13 for the pattern P, and 13.fwdarw.14.fwdarw.14 for the pattern Q.
In the 125 .mu.sec period, the number of bits of the DS3 payload is 5592.times.84/85=5526.211 . . . The number of bits in the PLCP multiframes is 5524 when the trailer length is 13 nibbles, and 5528 when the trailer length is 14 nibbles. Therefore, the cycle of the PLCP multiframe is fast on the cycle of 125 .mu.sec when the PLCP multiframe pattern is P, and is behind on the cycle of 125 .mu.sec when the PLCP multiframe pattern is Q.
Conventionally, the cycle of a transmitted PLCP frame is monitored, and the phase of the extracted clock is compared with the phase of the 8 KHz clock obtained by dividing 44.736 MHz. If the phase of the PLCP multiframe to be transmitted is forward, the trailer pattern is switched to P. If it is behind, the trailer pattern is switched to Q. Thus, the transmission cycle of the PLCP multiframe is adjusted properly.
FIGS. 910 and 911 are timing charts showing the circuit configuration and the operation for realizing the above listed functions.
A PLCP frame cycle monitoring unit 7 monitors the transmission cycle of the PLCP frames to be transmitted from a selector 3 to output a phase comparison pulse S for every third PLCP frame. A dividing unit 6 generates 8 KHz clock by dividing 44.736 MHz clock by 5,592 generated by a clock generating unit 5. A phase comparing unit 8 compares the phase comparison pulse S with the phase of the 8 KHz clock, and outputs a pattern switch signal C as a value of 1 when the phase comparison pulse S is behind and a value of 0 when is forward.
The selector 3 selects input A1 and A2 according to the pattern switch signal C. That is, the selector 3 selects the pattern P when the pattern switch signal C indicates 0 and selects the pattern Q when it indicates 1.
The PLCP frame generating units 1 and 2 for the patterns P and Q store an ATM cell or an L2-PDU cell in the PLCP payload and add a PLCP header and trailer to assemble a PLCP frame.
The pattern P PLCP frame generating unit 1 adds a trailer for indicating the number of nibbles 13, 14, and 13 on three cycles. The pattern Q PLCP frame generating unit 2 adds a trailer for indicating the number of nibbles 13, 14, and 14 on three cycles.
The DS3 interface unit 4 inserts a PLCP frame into the DS3 payload and adds a DS3 header to assemble and transmit a DS3 frame.
However, the above described conventional technology selects a trailer pattern according to the phase comparison result, and the transmission order of the pattern P and Q is not fixed.
As a result, there arises a problem that the complicated operations generate a complicated circuit.
Additionally, there is a problem of a large deviation of transmission timing.
The following functions are required to realize the multicasting capabilities (point-to-multipoint connection) in the ATM switch.
The efficiency in use of the resources as a switch is higher when cells are copied at a point nearer to the exit of the exchange station. The copied cells are distributed to each subscriber. The cells distributed to each subscriber has different VPI/VCIs. That is, the VPI/VCI depends on the destination subscriber. The number of bits of the VPI/VCI is equal to or larger than 22 bits. Simply converting the large number of bits undesirably results in large-scale hardware.
The ATM switch exchange cells in a self-routing system. If a large-capacity system performs a self-routing process, the efficiency of the switch is higher when the multicasting capabilities are supported in the switch. Thus, the entire system can be smaller in size with the cost reduced.
The services supported in the B-ISDN should include a large number of point-to-multipoint connection services as well as multicasting capabilities. To reduce the scale of the entire switch, the multicasting capabilities added to realize the point-to-multipoint connection should be minimized for smaller scale and cost. Furthermore, the future extension of the multicasting capabilities should be considered.
In the point-to-multipoint connection, such information as specifies the number of copied cells and the destination of each of the copied cells is required. The information is normally set as tag information added to the cell when it is input to the exchange station. However, since the amount of the above described information is not small, the tag information occupies about 10 bytes. Adding such tag information to a cell makes the entire cell length longer than in the exchange station. That is, when the tag information is longer, the ratio of the actual data to the entire cell becomes smaller, thereby lowering the throughput.
FIG. 912 shows the configuration of the form of the conventional multicasting capabilities. In FIG. 912, a source terminal 1 multicast-transfers data to destination terminals 4-1-4-5 through an ATM switch 2.
Line 3 connects the source terminal 1 with the ATM switch 2. The line 3 can multiplex and transmit a plurality of calls (paths). The ATM switch 2 is also connected to the destination terminals 4-1-4-5 through a subscriber line capable of multiplexing and transmitting data. In the ATM switch 2, a virtual path is set according to the destination information written in the cell transmitted by the source terminal 1. In the example shown in FIG. 912, virtual paths 5-1-5-5 are set as paths for transferring cells to the destination terminals 4-1-4-5.
In the above described multicasting transfer, cells are copied for the destination terminals in the source terminal 1 and transferred through the paths set between the source terminal 1 and the destination terminals 4-1-4-5. At this time, 5 channels are multiplexed in the line 3 to transfer cells to the destination terminals 4-1-4-5. That is, the bands of 5 channels are occupied.
Thus, since N paths are set between the source terminal and destination terminal when 1:N multicast transfer is made according to the conventional method shown in FIG. 912, the resources for the line 3 and ATM switch 2 have been used more than necessary and the load on the source terminal 1 has been heavy.
It is expected that the demand for dynamic images will greatly increase. For example, members of companies in the distance have a lot of opportunities to have things settled through conferences over telephone using dynamic images. These services not only satisfy individual subscribers but also promote business smoothly regardless of geographical disadvantages.
Nevertheless, these services have not been sufficiently offered. That is, the 1:1 communications are more popular than the private line services in the broadband communications network, and the method of controlling the multi-terminal connection, for example, a three-subscriber communications has not been put to practical use.
Described below is the problem related to the process performed in the event of a failure on a device in the exchange station which processes a transmission line.
With the ATM switch, a communications line system device in the exchange station processes a number of virtual lines (hereinafter referred to simply as lines) specified by individual VPI/VCIs. When a failure occurs on a communications line system device, how to handle the lines processed by the device is very important in maintaining the quality of the communications.
When a failure occurs on a communications line system device in the exchange station, a call connected through the line processed by the device is compulsorily terminated by a compulsory release process activated by the fault monitor process for the entire system. Therefore, the subscribers have the problem that the communications may be suddenly terminated.
The conventional systems have not provided the mechanism of managing the line processed by the communications line system device.
Described below is the problem relating to the process performed when a failure is detected on the line.
When a line failure is detected on a single-structured, not duplex, ATM switch, the transmission information such as subscriber information, billing information, traffic information, performance information, etc. is saved by a line switch process in physical line units using a reserved line, etc. conventionally.
Practically, if a failure is detected on one physical line when a remote concentrator 1 and an ATM switch 2 are connected through a plurality of physical lines as shown in FIG. 913, then the faulty band or an idle band for other lines are not used, but the state of the faulty line is assigned to a new alternate line such as a spare line, etc.
Therefore, even though large idle bands exist in other lines, they are not utilized effectively, thereby lowering the use rate of the lines.
To perform a line switch process in physical line units, it is necessary either to reserve sufficient spare lines or to duplex each of the physical lines. As a result, the communications may cost high.
It is also necessary to duplex the intra-station device such as a communications system device, etc. in the exchange station to maintain the reliability of the communications. If a failure occurs on the intra-station device of the active system, then various communications control data are transferred to the intra-station device of a standby system to stop the operation of the intra-station device which has been a device in the active system and start the operation of the intra-station device which has been an intra-station device of the standby system.
In this case, various communications control data set in the intra-station device of the active system have been conventionally transferred to the intra-station device of a standby system by a processor controlling the intra-station device. However, since the amount of the various communications control data is large for the ATM switch, etc., a long time is required by the processor to transfer the data from the intra-station device of an active system to the intra-station device of a standby system, thereby disadvantageously affecting the reliability of the exchange station when a failure occurs on the exchange station.